Purpose: We recently reported that children with acute leukemias who show increasing mixed chimerism (MC) after allogeneic stem-cell transplantation have a significantly enhanced risk of relapse. Here we present the results of a prospective multicenter study to investigate (1) whether relapse of acute lymphoblastic leukemia (ALL) can be determined in advance by serial analysis of chimerism, and (2) if outcome can be influenced by withdrawal of immunosuppression and/or by low-dose donor lymphocyte infusion when increasing MC is detected.
Patients and methods: Serial and quantitative analysis of chimerism was performed using a fluorescent-based short-tandem-repeat-polymerase chain reaction in 163 children with ALL.
Results: One hundred one patients revealed complete chimerism (CC) or low-level MC (CC/low-level MC); increasing MC was found in 46 patients; and decreasing MC, in 16 patients. Relapse was significantly more frequent in patients with increasing MC (26 of 46) than in patients with CC/low-level MC (eight of 101) or in patients with decreasing MC (0 of 16; P <.0001). The probability of 3-year event-free survival (EFS) was 54% for all patients, 66% for patients with CC/low-level MC (n = 101), 66% for patients with decreasing MC (n = 16), and 23% for patients with increasing MC (n = 46; P <.0001). Of the 46 patients with increasing MC, 31 received immunotherapy. This group had a significantly higher 3-year EFS estimate (37%) than the 15 patients who did not receive immunotherapy (0%; P <.001).
Conclusion: Serial analysis of chimerism reliably identifies patients at highest risk to relapse. The 3-year EFS of patients with increasing MC without immunotherapy was 0%, by which overt relapse could be prevented in a considerable group of patients.